Whole genome disease screen - in under 50 hours - for newborn babies

07 October 2012

Interview with

Dr Stephen Kingsmore, Children’s Mercy Hospital in Kansas City

Chris:: A new diagnostic tests for new-born babies has been developed.  It can screen for more than 3,500 genetic disorders within 50 hours, a very big improvement when you compare it with the currently-used methods.  DNA HelixThree per cent of newborns will have a genetic disorder and for some of these problems, the earlier that you start treating, the better the overall prognosis.  With us to discuss this new test, which has been developed with the help of a company called "Illumina" - who have their European base in Essex - is Dr. Stephen Kingsmore.  He's from the Children's Mercy Hospital in Kansas City and he's with us.  Hello, Stephen.

Stephen::  Hello.

Chris::   So, what was the problem you're actually trying to solve with this?

Stephen::  Let me give you an example before and after this test.  So baby Stephanie was born in March 2012.  She was a bit small at first, but pretty normal.  However, 1 hour after she was born, she started to have seizures, epileptic fits and so she was transferred to an intensive care unit.  And for 5 weeks, pretty much everything was attempted to stop these seizures, but without any effect.  In that period of time, her parents had tremendous anguish.  They had feelings of guilt and they had false hope that their baby is going to be saved; and all the while, nobody could tell them exactly what was going on.  Now eventually, after 5 weeks, the doctors got together with the parents.  They said, "We think this is hopeless."  And the parents after some thought decided to withdraw support and baby Stephanie died that day.  Now let's contrast that with after the test.  So now, let's say baby Stephanie was going to be born in December of this year, she'd have exactly the same first day of life, but 2 days later, two days later she would've received, or her parents would've received, a definitive diagnosis that she had a mutation in a gene called BRAT1 and that, tragically, there was no way to save her life.  What that would have meant for her parents is that, first of all, they wouldn't have been guilty.  This was not because mum had a glass of wine during pregnancy.  They would've known that it was hopeless.  It would've had a clear.  It would've known their risk of having another baby with the same thing and they would've been able to bond with their baby with the baby out of the intensive care unit, be able to say their goodbyes, bring in their pastor or priest, bring in their grandparents, and there would've been a sense of parental empowerment, and proper goodbyes.  So that's the difference before and after this test was developed.

Chris::  Sounds wonderful.  How does it actually work?

Stephen::  Well, it's pretty simple.  We decode the entire human genome using a very speedy sequencer.  So that's 3.1 billion letters - that's the equivalent of a manuscript 300 feet high - and we do that in each baby.  We do that in a day and then it takes another day for us to take that information, process it and link it to whether or not the baby had one of the 3,500 genetic diseases.

Chris::  First of all, the human genome project, which sequenced one human, took years to do it and you're doing that in a day.  How on Earth are you doing that?

Stephen::  This was a technology that was developed at Oxford University about almost a decade ago, but it's been getting better and better.  And we're at the point now where we can decode a baby's entire genome in roughly a day.  I mean, the scale has gone remarkably so that we have, literally, billions of chemical reactions occurring that we can monitor in real time and churn out hundreds of billions of very short fragments of the human genome code, and then we put these together to get the whole piece.  It's like making a jigsaw puzzle with, say, 600 billion pieces.

Chris::  But how do you know which of the changes in the DNA are the crucial ones that are giving that child its problem?

Stephen::  It's been a huge problem, because only a tiny fraction of doctors, paediatric geneticists, actually know which test to order.  There are 3500 to choose from and most physicians wouldn't have heard of most of these diseases.  And so, we built some software that allows a regular paediatrician to plug in what's going on in the baby.  In Stephanie, she was small and she was having seizures.  So the doctor clicks those two boxes and a computer program matches that information to just the points in the genome where there are 89 genes that can cause this in a baby.  And in that way, instead of having to scour the whole genome, we're able to focus in on the specific areas where, if there is a genetic cause, that's where it would be pinpointed.

Chris::  And can it learn to recognise new disorders as more discovered because the power of this is obviously the rapid diagnosis.  The downside is that it only answers the question you ask it, 'has the child got one of these conditions we know about'?

Stephen::  That's a good question, and I don't have an easy answer; but if you look at the situation today where it takes 2 months to look at a single gene, a single point in the genome for a single disease, and you compare that with the ability in 2 days to look at 3,500, this is still a phenomenal advance.  And building it out so that it sort of self-learns is definitely feasible, but not straightforward.

Chris::  And is this widely available?  Will the average hospital, maternity or maybe specialist unit have this technology available to it in the near future?

Stephen::  Well, I should say that our paper is a proof of concept.  We only looked at about half a dozen babies and so, this really is just a first report.  A lot more needs to be done to understand what are the right places to put this in intensive care units.  It's expensive.  It costs over 8,000 pounds to test a single baby and the sort of infrastructure is not something you find commonly in hospitals.  But we hope by the end of this year - so Illumina are shipping us our new genome analyser - but the work we did in the manuscript, we actually had to get the DNA from each baby and then courier it across the Atlantic to Essex, where the DNA sequence was generated, and then they loaded that on a hard drive and shipped it back across the Atlantic to Kansas City where we crunched the information.  So, our goal, by December, is to have our own instrument in our own hospital and start to build experience.  We really think that maybe next year, it will be possible for some places in the United States, and probably also in Great Britain, to be referral centres.  You don't necessarily have to have the baby in your intensive care unit.  You can certainly ship the DNA to a central location where this service could be offered.

Chris::  Super!  Well, thank you for joining us to tell us about it, Stephen.  It sounds very exciting.  That was Dr. Stephen Kingsmore.  He's from the Children's Mercy Hospital in Kansas City and he published that research this week in the journal Science Translational Medicine.

Add a comment